Application of a maximum likelihood algorithm to ultrasound modulated optical tomography

Area of Science:

• Biomedical optics
• Medical imaging
• Ultrasound technology

Background:

• Pulsed ultrasound modulated optical tomography (USMOT) uses ultrasound pulses to scan samples and modulate scattered light.
• Detecting modulated signals allows for 1D imaging along the ultrasound axis.
• Existing methods require robust models for accurate image reconstruction.

Purpose of the Study:

• To develop and validate a maximum likelihood (ML) image reconstruction method for pulsed USMOT signals.
• To investigate the theoretical and experimental performance of the ML technique for USMOT.
• To assess the capability of the ML method in retrieving spatially varying optical properties.

Main Methods:

• A simple convolution model was developed for the detected USMOT signal.
• A maximum likelihood (ML) method was established based on the convolution model.
• Simulations and experiments with tissue phantoms were used for validation.

Main Results:

• The ML method successfully inverted USMOT data to retrieve sample properties along the ultrasound axis.
• Simulations showed the ML method is effective even with a signal-to-noise ratio (SNR) near unity.
• Experimental results with tissue phantoms yielded an axial resolution of 160 μm and lateral resolution of 600 μm.

Conclusions:

• The developed ML method is a viable tool for pulsed USMOT image reconstruction.
• The technique can provide signals proportional to optical properties.
• High resolutions were achieved, demonstrating the potential for detailed imaging.